How to Write AI Prompts for Sales Copy That Actually Convert

Most AI copy is garbage.
You ask for a sales page, you get generic fluff that wouldn’t convert a starving man to a cheeseburger. That’s because you’re treating the AI like a magic button instead of a junior copywriter who needs a strict brief.
I’ve spent years running campaigns where every word carries a cost. If the copy doesn’t pull, you don’t get paid. Simple as that.
The difference between a dud and a winner? Structure. I use two frameworks to keep the machine on a leash: PAS and RIGS. Here is how I run them.
Stop Guessing. Start Prompting.
Sales copy has one job: get the reader to move. To do that, you need to hit their pain points and give them a clear path to the solution. When you feed an AI a vague request, it guesses. And it guesses wrong.
A good prompt acts like a creative brief. It’s the set of rails the AI has to run on.
For agency owners, I don’t just fire off prompts into the void. I organize these workflows inside GoHighLevel. Storing your best prompts right next to your client assets? That’s how you actually scale.
Speed is good, but conversion is the only metric that matters. Let’s get into the frameworks.
1. The PAS Framework (Problem, Agitate, Solution)
This is the bread and butter of direct response. It’s older than the internet, and it works better than any modern “creative” AI output.
You define the problem. You turn the knife (agitate). Then you offer the fix.
Here is my go-to template:
“Write a sales email for [product]. First, describe the specific pain [customer segment] feels when [specific pain point]. Then, agitate that problem by highlighting the hidden costs of doing nothing. Finally, present my product as the solution, focusing on [primary benefit].”
PAS is for when you need to pull on heartstrings and drive an emotional decision. If you’re building out email sequences to nurture these leads, automate the delivery here. It’s how you stay top-of-mind without touching a keyboard.
2. The RIGS Method (Role, Instruction, Guardrails, Specifics)
When I need precision—like for a cold outreach campaign or a direct-response ad—I use RIGS. It removes the guesswork.
Role: Who is the AI today? (e.g., “You are a direct-response copywriter.”)
Instruction: What is the output? (e.g., “Write a 150-word LinkedIn ad.”)
Guardrails: What are the hard “no’s”? (e.g., “No jargon. No ‘game-changer’ fluff.”)
Specifics: What is the data? (e.g., “Product: Automated CRM. Target: Small business owners.”)
Simple. Effective. No fluff.
|
Framework |
Best Used For |
The Goal |
|---|---|---|
|
PAS |
Landing pages / Emails |
Emotional Trigger |
|
RIGS |
Ads / Cold Outreach |
Precise Compliance |

The Operator’s Reality Check
You want to know the biggest mistake I see? Skipping the “Role” assignment. If you don’t tell the AI who it is, it defaults to a boring, corporate tone that nobody wants to read.
Also, stop expecting publish-ready copy on the first try. AI is a draft engine. I spend 20 minutes prompting, and another 10 refining. That’s the work.
If you’re still doing this manually, you’re bleeding time. Use a platform that ties your copy pipeline to your lead management. Grab a trial of GoHighLevel and stop playing small.
Common Questions
Which framework is better?
Neither. PAS is for feelings; RIGS is for facts. Test both against your audience and let the conversion data tell you the winner.
Do I need a paid prompt library?
No. If you know the two frameworks above, you have everything you need. Save your money for ad spend.
Can I automate this fully?
You can automate the delivery, but never the final edit. Keep your eyes on the copy before it goes live. You’re the human. You’re the one who knows the brand voice.

Look, stop serving the algorithm. Build your own systems, use these frameworks to speed up your output, and stop relying on “magic” prompts that don’t exist.
